Because the combination of different legislator-level covariates (e.g., state, age, migration background) makes it possible to identify individual legislators 
and thus find out about their response behavior toward constituents, which is potentially sensitive information, we uploaded our replication data 
in the following form: We created two separate data sets that are not linkable through a common ID (data_experiment_outcomes_csv.csv and data_experiment_covariates_csv.csv). 
The first data set contains the outcome variables measuring the legislators’ response behavior and the treatment variables. The second data set contains the 
pre-treatment covariates and the treatment variables.

This allows the replication of our main models that estimate the treatment effects and our tables summarizing descriptive statistics and treatment balance 
on covariates without undermining the anonymity of our experimental subjects. While a replication of our models measuring the robustness of the treatment effects 
by including pre-treatment covariates is not possible with the publicly available replication data, please note that our estimates do not change substantially 
after the inclusion of those covariates.

For the reasons described above, the following tables and figures cannot be replicated, even in parts, with the replication data we uploaded: 
 - Figure A3
 - Table A11
 - Table A14